The structure of your merge project determines if all data is put in one (large) output document or if each row is put into a new output document.
To create one output document per row in the spreadsheet/database, the output document node must be a child of the data node.
Because the data node will repeat all its child nodes, it will create a new output document for each row it finds.